Reagent-Free Hyperspectral Diagnosis of SARS-CoV-2 Infection in Saliva Samples

Background: Rapid, reagent-free pathogen-agnostic diagnostics that can be performed at the point of need are vital for preparedness against future outbreaks. Yet, many current strategies (polymerase chain reaction, lateral flow immunoassays) are pathogen-specific and require reagents; whereas others such as sequencing-based methods; while agnostic, are not (as yet) conducive for use at the point of need. Herein, we present hyperspectral sensing as an opportunity to overcome these barriers, realizing truly agnostic reagent-free diagnostics. This approach can identify both pathogen and host signatures, without complex logistical considerations, in complex clinical samples. The spectral signature of biomolecules across multiple wavelength regimes provides rich biochemical information, which, coupled with machine learning, can facilitate expedited diagnosis of disease states, the feasibility of which is demonstrated here. Innovation: First, we present ProSpectral™ V1, a novel, miniaturized (~8 lbs) hyperspectral platform with ultra-high (2-5 nm full-width, half-max, i.e., FWHM) spectral resolution that incorporates two mini-spectrometers (visual and near-infrared). This engineering innovation has enabled reagent-free biosensing for the first time. To enable expedient outcomes, we developed state-of-the-art machine learning algorithms for near real-time analysis of multi-wavelength spectral signatures in complex samples. Taken together, these innovations enable near-field ready, reagent-free, expedient agnostic diagnostics in complex clinical samples. Herein, we demonstrate the feasibility of this synergy of ProSpectral™ V1 with machine learning to accurately identify SARS-CoV-2 infection status in double-blinded saliva samples in real-time (3 seconds/measurement). The infection status of the samples was validated with the CDC-approved polymerase-chain reaction (PCR). We report accuracies comparable to first-in-class PCR tests. Further, we provide preliminary support that this signal is specific to SARS-CoV-2, and not associated with other respiratory conditions.     Interpretation: Preparedness against unanticipated pathogens and democratization of diagnostics requires moving away from technologies that demand specific reagents; and relying on intrinsic biochemical properties that can, theoretically, inform on all pathologies. Integration of hyperspectral sensors and in-line machine learning analytics, as reported here, shows the feasibility of such diagnostics. If realized to full potential, the ProSpectral™ V1 platform can enable agnostic diagnostics, thereby improving situational awareness and decision-making at the point of need; especially in resource-limited settings – enabling the distribution of newly developed tests for emerging pathogens with only a simple software update. Funding: This work was supported by the Defense Threat Reduction Agency “Multiwavelength Spectroscopy of Innate Immunity, (PI Mukundan, PM Dr. Wallace) at the Lawrence Berkeley National Laboratory. Lawrence Berkeley National Laboratory is a multi-program national laboratory operated by the University of California for the DOE under contract DE AC02- 05CH11231. This work was also supported by the U.S. Department of Energy through the Los Alamos National Laboratory. Los Alamos National Laboratory is operated by Triad National Security, LLC, for the National Nuclear Security Administration of the U.S. Department of Energy (Contract No. 89233218CNA000001). The research presented in this article was supported by the Pathfinder/Technology Evaluation and Demonstration program of Los Alamos National Laboratory. Declaration of Interest: AWY, SG, and MK were full-time employees of Pattern Computer Inc. at the time of the study.